In-situ Test based methods:

Soil liquefaction potential can be determined by using in-situ tests such as: (i) standard penetration test (SPT) (ii) cone penetration test (CPT) (iii) shear wave velocity (Vs)

measurement (iv)Becker penetration test (BPT).

Due to difficulties in obtaining high quality undisturbed samples and subsequent high quality laboratory testing of granular soils, use of in-situ tests along with case histories- calibrated empirical relationships are generally resorted by the geotechnical engineers for the assessment of liquefaction potential of soils. The simplified procedure pioneered by Seed and Idris (1971) mostly depend on a boundary curve, which presents a limit state and separates liquefaction cases from the non-liquefaction cases basing on field observations of soil in earthquakes at the sites where in situ data are available. The boundary is usually drawn conservatively such that all cases in which liquefaction has been observed lie above it. In this approach the CSR is usually used as earthquake loading parameters and the cyclic resistance ratio (CRR) is represented by in-situ test parameters that reflect the density and pore pressure generation properties of soil. Out of the various in-situ methods as mentioned above SPT and CPT-based methods are widely used for liquefaction susceptibility analysis of soil.

SPT-based method

It is the most widely used methods among the available in-situ test methods as discussed above for evaluation of resistance of soil against the occurrence of liquefaction. Whitman (1971) first proposed to use liquefaction case histories to characterize liquefaction resistance in terms of measured in situ test parameters. Seed and Idriss (1971) did a pioneer work in developing a simplified empirical model, using laboratory tests and post liquefaction field observations in earthquakes, which presents a limit state function separating liquefied cases from the non-liquefied cases on the basis of SPT data. Seed et al. (1983) extended their

20

previous work in developing a modified model in which used CSR (τav/σv’) instead of peak

ground acceleration (amax) as a measure of seismic action and overburden pressure corrected

SPT value (N1) instead of relative density (Dr) as the site parameter representing its

resistance to liquefaction. However, it has been addressed by many researchers that the SPT has been conventionally conducted by using different kinds of hammers in different parts of the world, with different energy delivery systems, which also have varying degrees of efficiency. Moreover, the borehole diameters and the sampling techniques also differ significantly, which in turn cause a large variability in the measured values depending on the combinations of actual test procedures and equipment used.

Seed et al. (1985) expressed the measured penetration resistance (Nm) in terms of N1,60

where the driving energy in the drill rod is considered to be 60% of the free fall energy and correction for overburden effect is applied. Liquefaction resistance curves for sands with different fines contents are proposed, which is considered to be more reliable than the previous curves expressed in terms of mean grain size. Cyclic stress ratio, CSR, as proposed by Seed and Idriss (1971) and its subsequent modifications in Seed et. al.(1983), Seed et al.(1985), Youd et al. (2001), is defined as the average cyclic shear stress, τav, developed on

the horizontal surface of soil layers due to vertically propagating shear waves normalized by the initial vertical effective stress, σ′v, to incorporate the increase in shear strength due to

increase in effective stress and is presented as follows:

d v v v av _r g a CSR max_' ' 0.65       (2.4)

where σv’ = effective vertical stress at the depth under consideration. The value of CSR is

corrected to an earthquake magnitude of 7.5, using the magnitude correction proposed by Seed et al. (1985). Seed et al.(1985) proposed a standard blow count N60as given below:





60 N ER

N  m (2.5)

where ER= percentage of the theoretical free-fall energy (i.e., estimated rod energy ratio expressed in percentage); and Nm= measured SPT blow count corresponding to the ER. The

21

value of N60is corrected to an effective stress of 100 kPa. Thus, the overburden stress and

energy corrected SPT value, N1,60 is obtained by using the following relation:

60 60

,

1 C N

N  _N  (2.6)

where CN is the effective stress correction factor and is calculated from the following

relation:





where, Pa = 1atm of pressure in the same units used for σ′v. Fig. 2.2 is a graph of calculating

CSR and corresponding N1, 60 data from sites where liquefaction was or was not observed

following past earthquakes with magnitudes of approximately 7.5. Liquefaction and non- liquefaction data were separated by Cyclic Resistance Ratio (CRR) curves. Curves were developed for granular soils with the fines content of 5% or less, 15%, and 35%. Fig. 2.2 is only applicable for magnitude of 7.5 earthquakes.

Juang et al. (2000) proposed an artificial neural network (ANN) -based CRR model based on SPT dataset and used Bayesian mapping function approach to relate factor of safety against the occurrence of liquefaction, Fs with probability of occurrence of liquefaction, PL. Youd et

al. (2001) published a summary paper of 1996 and 1998, NCEER workshop in which the updates and augmentations to the original “simplified procedure” of Seed and Idriss (1971); Seed et al.1983; and Seed et al (1985) for evaluation of liquefaction potential, are recommended using SPT-based methods and is still followed as the current state of the art on the subject of liquefaction potential evaluation. Cetin (2000) and Cetin et al. (2004) proposed new correlations for assessment of liquefaction triggering in soil.

22

These correlations are developed on the basis of an expanded and reassessed post liquefaction SPT database after making screening of field data case histories on a quality/uncertainty basis, incorporating improved knowledge and understanding of factors affecting interpretation of SPT data, using improved understanding of factors affecting site specific earthquake ground motion, implementing improved methods for assessment of in situ CSR and using higher order probabilistic tools, Bayesian updating technique. The resulting correlations reduce the uncertainty associated with the liquefaction potential evaluation with respect to the existing models and also resolve controversial issues like magnitude-correlated duration weighting factors, adjustment of fines content and corrections for overburden stress in the context of assessment of CSR. Idriss and Boulanger (2004) and Idriss and Boulanger (2006) re-examined the existing semi-empirical procedures for evaluating the liquefaction potential of saturated cohesion-less soils during earthquakes and recommended revised correlations for use in practice. In this paper the authors discussed about the parameters, which contribute to the CSR formulation like stress reduction factor, earthquake magnitude scaling factor, overburden correction factor, and also the overburden

Fig. 2.2 SPT –based limit state boundary curves for Magnitude 7.5 earthquakes with data from liquefaction case histories (Modified from Youd et al. 2001)

23

normalization factor for penetration resistances and presented the modified relations for these parameters.

CPT-based method

Although, the above SPT-based method remains an important tool for evaluating liquefaction resistance, it has some drawbacks, primarily due to the variable nature of the SPT (Robertson and Campanella, 1985; Skempton, 1986), nowadays the cone penetration test (CPT) is becoming more acceptable as it is consistent, repeatable and able to identify a continuous soil profile. Thus, CPT is being used as a valuable tool for assessing various soil properties, including liquefaction potential of soil. A typical CPT involves pushing a 35.7mm diameter conical penetrometer into the ground at a standard rate of 2cm/sec, while electronic transducers record (generally at 2cm or 5cm intervals) the force on the conical tip, the drag force on a short sleeve section behind the tip, pore water pressure behind the tip (or sometimes at other locations). The tip force is divided by the cross sectional area of the penetrometer to determine the tip resistance, qc and the sleeve drag force divided by the

sleeve surface area to determine the sleeve friction, fs. The main advantages of the CPT are

that it provides a continuous record of penetration resistance and is less vulnerable to operator error than the SPT. The main disadvantages of the CPT are the difficulty in penetrating layers that have gravels or very high penetration resistance and need to perform companion borings or soundings to obtain actual soil samples.

Zhou (1980) first published liquefaction correlation directly based on case history CPT database of the 1978 Tangshan earthquake. He presented the critical value of cone penetration resistance separating liquefiable from non-liquefiable conditions to a depth of 15m. Seed and Idriss (1981) as well as Douglas et al. (1981) proposed the use of correlations between the SPT and CPT to convert the available SPT-based charts for use with the CPT data. Robertson and Campanella (1985) developed a CPT- based method for evaluation of liquefaction potential, which is a conversion from SPT-based method using empirical correlation of SPT-CPT data and follows the same stress-based approach of Seed and Idriss (1971). This method has been revised and updated by many researchers (Seed and de-Alba 1986; Shibata and Teparaksa 1988; Stark and Olson, 1995; Suzuki et al. 1995; Olsen 1997,

24

Robertson and Wride 1998). Most of the CPT based simplified methods are presented in a chart that defines the limit state function (i.e., a boundary curve) separating the liquefied and non-liquefied cases in a plot of the cyclic resistance ratio (CRR) versus corrected CPT tip resistance (QC). These methods also need the knowledge of mean particle size (D50) and

fines content (FC) which cannot be obtained from CPT measurements alone. For determining D50 and FC additional boreholes are required for collecting samples. Ishihara

(1993) suggested that in case of liquefaction resistance evaluated by using CPT value for silty sands (>5% fines), the effects of fines could be estimated by adding some tip resistance increments to the measured tip resistance to obtain an equivalent clean sand tip resistance. For evaluating liquefaction potential only from CPT measurements, Olsen (1997) developed a CRR model using the parameters: qc, σ′v and friction ratio (Rf). Robertson and Wride

(1998) proposed a separate method using soil behaviour type index, Ic, which was

recommended for use by the 1998, National Center for Earthquake Engineering Research (NCEER) workshop and is also presented in the summary paper of Youd et al. (2001).Fig. 2.3is used to determine the CRR for clean sands [i.e., fines content (FC) ≤5%] from CPT data. This chart (i.e., Fig. 2.3) is valid for the magnitude 7.5 earthquake only.

As per Juang et al. (1999a), Robertson and Wride method and Olsen method are found to be quite comparable. Juang et al. (2003) also developed an ANN-based simplified method using soil type index (Ic) for evaluation of CRR of soil using post liquefaction CPT database

and also used Bayesian mapping function approach to relate Fs with PL. Moss (2003) and

Moss et al. (2005) presented a CPT-based probabilistic model for evaluation of liquefaction potential using reliability approach and a Bayesian updating technique. Juang et al. (2006) used first order reliability method (FORM) for probabilistic assessment of soil liquefaction potential.

25

Shear wave velocity (Vs)-based methods

The use of shear wave velocity (Vs) as a in-situ test index of liquefaction resistance of soil is

very well accepted because both Vs and CRR are similar, but not proportional, influenced by

void ratio, effective confining stresses, stress history, and geologic age. The followings are the main advantages of using Vs for evaluation of liquefaction potential: (i) Vs measurements

are possible in soils that are difficult to penetrate with SPT and CPT or difficult to extract undisturbed samples, such as sandy and gravelly soils, and at sites where borings or soundings may not be permitted; (ii) Vs is a basic mechanical property of soil materials,

directly related to small-strain shear modulus; and (iii) the small-strain shear modulus is a parameter required in analytical procedures for estimating dynamic soil response and soil- structure interaction analyses. But, the following disadvantages are also there when Vs is

used for liquefaction resistance evaluations: (i) seismic wave velocity measurements are made at small strains, whereas pore-water pressure build up and the liquefaction triggering are medium- to high-strain phenomena; (ii) seismic testing does not provide samples for

Fig.2.3 Curve recommended for calculation for CRR from CPT data (Reproduced from Robertson and Wride 1998).

26

classification of soils and identification of non-liquefiable soft clay-rich soils; and (iii) thin, low Vs strata may not be detected if the measurement interval is too large. Therefore, it is

preferred to drill sufficient boreholes and conduct in-situ tests (SPT or CPT) to detect and demarcate thin liquefiable strata, non-liquefiable clay-rich soils, and silty soils above the ground water table that might become liquefiable should the water table rise. Few VS-based

simplified methods (Dobry et al. 1981; Stokoe et al. 1988; Tokimatsu and Uchida 1990; Robertson et al. 1992; Kayen et al. 1992; Lodge 1994; Andrus and Stokoe 1997; Andrus and Stokoe 2000; Juang et al. 2000a; Juang et al. 2001; Andrus et al. 2003) have been developed and are in use. But as Vs method is of recent origin and has not been verified with the historical post liquefaction database, Vs – based method is not that popular like SPT and CPT–based method.

BPT-based methods

Liquefaction resistance of non-gravelly soils has been assessed mostly through SPT and CPT, with rare Vs measurements. Several investigators have employed large-diameter

penetrometers to overcome these difficulties; the Becker penetration test (BPT) in particular has become one of the more effective and widely used larger tools. The BPT was developed in Canada in the late 1950s and consists of a168-mm diameter, 3-m-long double-walled casing driven into the ground with a double-acting diesel-driven pile hammer. The hammer impacts are applied at the top of the casing and the penetration is continuous. The Becker penetration resistance is defined as the number of blows required to drive the casing through an increment of 300 mm. The BPT has not been standardized, and several different types of equipment and procedures have been used. There is currently very few liquefaction sites from which BPT data have been obtained. Thus the BPT cannot be directly correlated with field behaviour, but rather through estimating equivalent SPT Nm-values from BPT data and

then applying evaluation procedures based on the SPT. This indirect method introduces substantial additional uncertainty into the calculate CRR. But, very few BPT-based simplified methods (Harder and Seed 1986 and Youd et al. 2001) have been developed primarily as it is only suitable for gravelly soil.

27 2.3 METHODS OF ANALYSIS

The basic analysis criterion in liquefaction potential evaluation is to compare the resistance (CRR) of soil with the loading (CSR) effects. These liquefaction triggering analyses are carried out using the following three methods based on the importance of the project.

Deterministic method

Probabilistic method

Reliability-based probabilistic method

A brief description and literature pertaining to above methods are presented separately.